Hi All, Since the last stable point release (OSG-3.2.1) I've been merging various bug and build fixes to the OSG-3.2 branch. As it's a stable release the ABI has remained the same so you should be able to drop in directly against your applications.
http://www.openscenegraph.org/index.php/download-section/stable-releases My plan is to get a bit of feedback on build of OSG-3.2.2-rc1 on the major platforms and then go for a release in the first week of March, so we have next week to do testing and make any further 3.2.2 release candidates we need. Getting 3.2.2 out quickly will avoid too much overlap with making the OSG-3.4 which I will making in the second half March. As usual, I'll appreciate testing against all the platforms and applications you can muster :-) Thanks in advance, Robert. ChangeLog since 3.2.2 back on 4th July 2014. 2015-02-26 20:06 robert * AUTHORS.txt, ChangeLog, applications/osgversion/Contributors.cpp: Updated AUTHORS 2015-02-26 19:13 robert * src/osg/Image.cpp: From Bradley Baker Searles, "We had someone replace a DDS texture with a GL_RGB/GL_UNSIGNED_SHORT_5_6_6 image, which would trigger a crash in Image.cpp while flipping the mipmap chain. The code was trying to flip the mipmaps using the rowStep for the full-size image (we have "dds_flip" set in the osgDB::ReaderWriter::Options in the osgDB::Registry)." 2015-02-26 18:07 robert * src/osgPlugins/dds/ReaderWriterDDS.cpp: From Bradley Baker Searles, "Small patch attached to handle DDS flip when s=4. Currently it will omit flipping and generate a warning: "Vertical flip was skipped. Image dimensions have to be multiple of 4."." 2015-02-26 14:16 robert * src/osgPlugins/tiff/ReaderWriterTIFF.cpp: Fixed typo 2015-02-23 20:28 robert * src/osg/TransferFunction.cpp: Fixed image allocation operations. 2015-02-23 11:07 robert * src/osgViewer/StatsHandler.cpp: From Arjun Ramamurthy, fix for StatsHandler when being used in multiple view/windows 2015-02-18 10:23 robert * src/osgPlugins/dds/ReaderWriterDDS.cpp: From Sebastian Messerschmidt,"Attached you find a version of the ReaderWriterDDS with info outputs set to OSG_INFO instead of OSG_NOTICE. The old version spams the console for each loaded DDS file." 2015-01-06 15:10 robert * ChangeLog: Updated ChangeLog 2015-01-06 14:55 robert * src/osgVolume/Property.cpp: From Julien Valentin, added missing initializer 2015-01-06 14:41 robert * src/osgWrappers/serializers/osgVolume/VolumeTechnique.cpp: From Julien Valentin, typo fix 2014-12-22 10:35 robert * CMakeLists.txt: Added comment about using -DCMAKE_DISABLE_FINDPACKAGE_* command line. 2014-12-22 09:50 robert * include/osg/ref_ptr: Added doxygen comment for ref_ptr<>::release(). 2014-12-18 15:59 robert * CMakeLists.txt: From Mattias Helsing, "CMake have release 3.0 and 3.1 and we have some bad checks for cmake major version when settings cmake policies in CMakeLists.txt. This fixes it" 2014-12-18 11:37 robert * applications/osgversion/Contributors.cpp: Updated AUTHORS 2014-12-18 09:20 robert * PlatformSpecifics/Android/Application.mk.master.in: From Adrian Clark, "Recently I tried compiling version 3.x of OpenSceneGraph for Android, using a recent Android NDK, only to run into problems with missing dependencies when trying to link my android project against the generated libraries. After some exploration, I found the problem is because the APP_PLATFORM variable in Application.mk is set to a platform number which is no longer compatible with the Android NDK by the OSG CMake toolchain. This causes the Android to compile it using the most recent Android platform libraries which (for reasons I don't know) no longer provide functions which OSG relies (in particular rand and stat64). This bug affects the current SVN repos for OSG 3.3.1, OSG 3.3.0, OSG 3.2.1 and OSG 3.2.0 (and possibly earlier). I posted a longer discussion of what the problem is in this forum thread: http://forum.openscenegraph.org/viewtopic.php?p=62012#62012 But essentially the fix is to change one line in OSG/PlatformSpecifics/Android/Application.mk.master.in from: APP_PLATFORM := ${ANDROID_PLATFORM} To APP_PLATFORM := android-${ANDROID_PLATFORM} The attached file will fix the bug in the 3.2.0 - 3.3.1 SVN repos for OSG. " 2014-12-10 11:30 robert * src/osgPlugins/dae/daeRMaterials.cpp: From Farshid Lashkari, "The Collada loader would crash while processing textures on certain files. I've attached the fix." 2014-12-08 11:08 robert * src/osgUtil/CullVisitor.cpp: Fixed memory leak in RenderStageCache. 2014-11-26 17:07 robert * applications/present3D/Cluster.h: Fixed warnigns 2014-11-26 17:07 robert * applications/present3D/Cluster.cpp, applications/present3D/Cluster.h: Removed redundent line endings 2014-11-26 16:29 robert * src/osgPlugins/ffmpeg/CMakeLists.txt: Fixed deprecated warnings 2014-11-26 16:15 robert * src/osgPlugins/pdf/CMakeLists.txt: Fixed warning 2014-11-26 16:07 robert * examples/osgcluster/osgcluster.cpp: Fixed warnings 2014-11-26 16:06 robert * examples/osgcluster/osgcluster.cpp: Removed redudent spaces 2014-11-26 16:02 robert * src/osgPlugins/lwo/Object.cpp: Fixed warning 2014-11-24 15:24 robert * CMakeLists.txt, CMakeModules/FindAVFoundation.cmake, CMakeModules/FindQuickTime.cmake, src/osgViewer/PixelBufferCocoa.mm: From Jan Klimke, "i noticed, that there are a couple of additional flaws when building osg for Mac OS X 10.10 Yosemite. The mac os sdk version is recognized by the current CMAKE script as 10.1 instead of 10.10 since it cuts the version string from the 4th place. I introduced a more reliable version checking based on splitting the returned version code into MAJOR MINOR and PATCH parts and reassemble the OSG sdk version afterwards. I replaced the existing CMake code against the following (returning now version 10.10 as expected): # Determine the canonical name of the selected Platform SDK EXECUTE_PROCESS(COMMAND "/usr/bin/sw_vers" "-productVersion" OUTPUT_VARIABLE OSG_OSX_SDK_NAME OUTPUT_STRIP_TRAILING_WHITESPACE) STRING(REPLACE "." ";" MACOS_VERSION_LIST ${OSG_OSX_SDK_NAME}) LIST(GET MACOS_VERSION_LIST 0 MACOS_VERSION_MAJOR) LIST(GET MACOS_VERSION_LIST 1 MACOS_VERSION_MINOR) LIST(GET MACOS_VERSION_LIST 2 MACOS_VERSION_PATCH) SET(OSG_OSX_SDK_NAME "macosx${MACOS_VERSION_MAJOR}.${MACOS_VERSION_MINOR}") Also i added the check for the new Version to some more find scripts. Additionally the nil object in Objective C now seems to be equivalent with a null_ptr that cannot be passed as GLInt anymore. So i switched this in the PixelBufferCocoa.mm to pass a zero instead of nil. " 2014-11-24 14:10 robert * examples/osgsimplegl3/osgsimplegl3.cpp: Fixed handling of viewer's Camera 2014-11-21 20:16 robert * src/osgDB/CMakeLists.txt, src/osgDB/FileUtils.cpp: From Alberto Luaces,"the current code uses the preprocessor for generating the plugin path in a way that when CMAKE_INSTALL_PREFIX contains something along the lines of /usr/x86_64-linux-gnu/ it gets substituted as /usr/x86_64-1-gnu/ that is, the string is preprocessed again, thereby making changes to anything that matches any defined symbol, as "linux" in this example (https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=763816). Quoting that path directly in CMake scripts solves that problem. " 2014-11-21 10:21 robert * CMakeModules/CheckAtomicOps.cmake: From Björn Blissing, fix for ambiguous defines in Atomic.cpp when compiling with MinGW and GCC 2014-11-21 09:28 robert * applications/osgversion/Contributors.cpp: Copied Contributors file from svn/trunk. 2014-11-20 17:37 robert * include/osg/Texture: From Claus Steuer, "XCode 6, IOs 8.1 SDK Compile fix : There are some undefined texture formats when compiling osg for IOs 8.1 with XCode 6 and OpenGLES2 enabled." 2014-11-20 16:38 robert * src/osgPlugins/obj/ReaderWriterOBJ.cpp: From Farshid Lashkari, "The obj loader was overriding the existing database path list with the file path of the model, instead of prepending the file path to the path list. The latter seems to be more common behavior for most of the existing loader plugins. Also, the local options weren't actually being used when processing the scene graph for textures. I've attached the fix for both issues." 2014-11-20 10:52 robert * src/osgPlugins/Inventor/ConvertFromInventor.cpp: From Marc Helbling, "here is a trivial fix in the Inventor plugin. In one code path, the pointer validity is checked after dereferencing a pointer that can be null (image->valid()) instead of calling ref_ptr::valid (image.valid())." 2014-11-20 10:45 robert * src/osgUtil/MeshOptimizers.cpp: From Marc Helbling, "I've come across scenes that contains geometries with initialized but empty vertex arrays and primitives and that would make some optimizers crash. The submission therefore only contains a test on the size of the vertex array for the VertexCacheMissVisitor and the VertexAccessOrderVisitor visitors." 2014-11-20 09:55 robert * src/osgPlugins/3ds/ReaderWriter3DS.cpp: From Sukender, "Fix for 3DS reader, which may read wrong triangles. Actually, indices may suffer a 'short int' overflow, in two places." 2014-11-20 09:41 robert * src/osg/Texture2DArray.cpp: Fixed segfalt in Texture2DArray copy constructor where it would apply images to an uninitialzed vector. 2014-11-20 09:32 robert * src/osgPlugins/fbx/fbxMaterialToOsgStateSet.cpp, src/osgPlugins/fbx/fbxMaterialToOsgStateSet.h, src/osgPlugins/fbx/fbxRMesh.cpp: From Marc Helbling, "patch adding support for ambient material in the FBX plugin. It's a straightfoward adaptation of the emissive support and has been tested on a proprietary model." 2014-11-20 09:31 robert * CMakeLists.txt: Added default setting of build type to Release 2014-11-20 09:21 robert * src/osgPlugins/gif/ReaderWriterGIF.cpp: From Pjotr Svetachov, fix for build breakage with giflib 5.0. 2014-11-19 17:31 robert * applications/osgversion/Contributors.cpp: Fixed contributors names 2014-11-19 17:31 robert * src/osgPlugins/gif/ReaderWriterGIF.cpp: From Clement Boesch, "Fix remaining bit of Giflib5 usage" 2014-11-19 11:37 robert * src/osgWrappers/deprecated-dotosg/osg/ProxyNode.cpp: From Laurens Voerman, "I found a new way to crach the osgviewer: osgviewer "ProxyNode { FileNameList { cow.osgt } num_children 1 }".osgs The proxy node reader wrongly assumes options to be non NULL. fixed in attached zip: src\osgWrappers\deprecated-dotosg\osg\ProxyNode.cpp applies to both the 3.2 branch and svn trunk" 2014-11-04 16:29 robert * src/osgPlugins/ffmpeg/FFmpegDecoderAudio.cpp, src/osgPlugins/ffmpeg/FFmpegDecoderVideo.cpp: Build fixes for ffmpeg LIBAVCODEC_VERSION_MAJOR >= 56 2014-08-19 09:27 robert * src/osgPlugins/md2/ReaderWriterMD2.cpp: Fixed MD2 danglng pointer crash. 2014-08-08 16:42 robert * src/osgPlugins/txp/ReaderWriterTXP.cpp: From Ryan Kawicki, "There is an issue where the model insets of a terrex terrain are being removed during the loading process. The issue is described below. - the issue here is that the plugin is removing group nodes if that group node only has one child. becuase transforms are also group nodes, there were cases when the transform would have only one child under it and would cause it to remove the translation portion. this would cause all the vertex data to be loaded around the last matrix operation, which in our case was the origin (0,0,0). We work off of OSG 2.8.1 but see that this has not been addressed on latest yet. I've tested this against 2.8.1 and have cleanly applied it to my local repository off of latest." 2014-08-08 16:34 robert * src/osgUtil/RenderStage.cpp: From Tim George, "Currently there is a problem with using a camera with a viewport with a non 0 offset and also using an FBO. The problem is that only area made up of the viewports width and height is drawn based on an offset of 0,0 instead of using the viewports offset. It is caused by line 991 in RenderStage.cpp: Code: fbo_ext->glBlitFramebuffer( 0, 0, static_cast<GLint>(_viewport->width()), static_cast<GLint>(_viewport->height()), 0, 0, static_cast<GLint>(_viewport->width()), static_cast<GLint>(_viewport->height()), blitMask, GL_NEAREST); which is not taking into account the viewport x and y when performing the blit. It probably should be: Code: fbo_ext->glBlitFramebuffer( static_cast<GLint>(_viewport->x()), static_cast<GLint>(_viewport->y()), static_cast<GLint>(_viewport->width()) + static_cast<GLint>(_viewport->x()), static_cast<GLint>(_viewport->height()) + static_cast<GLint>(_viewport->y()), static_cast<GLint>(_viewport->x()), static_cast<GLint>(_viewport->y()), static_cast<GLint>(_viewport->width()) + static_cast<GLint>(_viewport->x()), static_cast<GLint>(_viewport->height()) + static_cast<GLint>(_viewport->y()), blitMask, GL_NEAREST); " Note from Robert Osfield, made small tweak to above on merge, changing the width+x to x+width to make it read more naturally. 2014-08-08 16:09 robert * src/osgUtil/MeshOptimizers.cpp: From Marc Helbling, "please find a fix for the vertex pretransform visitor (VertexAccessOrderVisitor). The issue with current code is that arrays are collected *before* duplicating shared arrays which leads to arrays that are correctly duplicated but that are not reordered. Also the submitted patch contains a small cleaning in GeometryArrayGathrer as the _useDrawElements variable is not used; it is only set in the GeometryArrayGathrer constructor and VertexAccessOrderVisitor already checks that primitives have indexed type." 2014-08-08 15:45 robert * include/osg/Node: Applied fix to Node::remove*Callback(NodeCallback*) inspired by fix from Glen Waldron that was applied to svn/trunk. 2014-07-29 15:47 robert * src/osgUtil/Optimizer.cpp: Added catch for NULL Transform pointers getting into the _transformMap. 2014-07-22 16:35 robert * include/osg/State, src/osg/State.cpp: Removed usage of _appliedProgramObjectSet as it's no longer used by OSG applications and was causing a threading crash. 2014-07-22 16:34 robert * CMakeLists.txt, include/osg/Version: Updated version to 3.2.2 2014-07-21 13:45 robert * src/osgText/Glyph.cpp: Added setting of glPixelStore before glTexSubImage2D call. 2014-07-21 13:43 robert * src/osgText/Glyph.cpp: I want to submit a bugfix for a crash occurring in osgText/Glyph.cpp if the scene tree contains (large) 2D textures from images with STRIDE. ============================================================================ #0 0x00007fffe8ea4350 in __memmove_ssse3 () from /lib64/libc.so.6 #1 0x00007fffe52ced76 in ?? () from /usr/lib64/libnvidia-glcore.so.310.44 #2 0x00007fffe52d8e86 in ?? () from /usr/lib64/libnvidia-glcore.so.310.44 #3 0x00007fffe53dd8be in ?? () from /usr/lib64/libnvidia-glcore.so.310.44 #4 0x00007fffe53c2643 in ?? () from /usr/lib64/libnvidia-glcore.so.310.44 #5 0x00007fffe53c7fdd in ?? () from /usr/lib64/libnvidia-glcore.so.310.44 #6 0x00007fffe53cbabf in ?? () from /usr/lib64/libnvidia-glcore.so.310.44 #7 0x00007fffe53cc1fa in ?? () from /usr/lib64/libnvidia-glcore.so.310.44 #8 0x00007ffff30092fd in osgText::GlyphTexture::apply (this=0x1bb8cf0, state= ...) at /d43/jaap/dev/jaapOSG/build/OpenSceneGraph3.3.1/src/osgText/Glyph.cpp:234 #9 0x00007ffff56c30b6 in osg::State::applyAttributeOnTexUnit (this=0x125f180, unit=0, attribute=0x1bb8cf0, as=...) at /d43/jaap/dev/jaapOSG/build/OpenSceneGraph3.3.1/include/osg/State:1713 #10 0x00007ffff56c2f3f in osg::State::applyTextureAttribute (this=0x125f180, unit=0, attribute=0x1bb8cf0) at /d43/jaap/dev/jaapOSG/build/OpenSceneGraph3.3.1/include/osg/State:411 #11 0x00007ffff30204da in osgText::Text::drawTextWithBackdrop (this=0x1baed70, state=..., colorMultiplier=...) ============================================================================== The crash disappears if I either (1) disable the use of images with stride in the (public) osgGeo-library, or (2) add the following bugfix to Glyph.cpp. This combination gives me the confidence that I understand where this problem originates from, without trying to understand the full OpenGL details. =============================================================================== @@ -221,7 +223,12 @@ imageData[i] = 0; } + glPixelStorei(GL_UNPACK_ALIGNMENT,1); + #if !defined(OSG_GLES1_AVAILABLE) && !defined(OSG_GLES2_AVAILABLE) + glPixelStorei(GL_UNPACK_ROW_LENGTH,getTextureWidth()); + #endif + // allocate the texture memory. glTexImage2D( GL_TEXTURE_2D, 0, GL_ALPHA, getTextureWidth(), getTextureHeight(), 0, ================================================================================ I have copied (and adapted) the added lines above from the same source file, where they were used in front of a similar call to glTexSubImage2D(.) around line 515."
_______________________________________________ osg-users mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
